AN EFFICIENT ALGORITHM TO FIND THE MLE OF PRIOR PROBABILITIES OF A MIXTURE IN PATTERN-RECOGNITION

A number of techniques have been proposed to determine the parameters which define the unknown components of a mixture in pattern recognitio n. The most common method is the maximum likelihood estimation (MLE). A direct ML approach requires solution to maximize the likelihood func tion of the unknown prior probabilities of classes in a mixture. This is a complicated multiparameter optimization problem The direct approa ch tends to be computationally complex and time consuming. In this stu dy, we use the concave property of the Kullback-Leibler information nu mber to derive a simple and accurate algorithm which can find the MLE of the prior probability of each class. The results of a Monte Carlo s imulation study with normal and exponential distributions are presente d to demonstrate the favorable prior estimation for the algorithm.